Message-ID: <03c53f05-3e2c-5b83-87a9-fbd91a36ba66@gmail.com>
Date: 2020-08-20T11:07:24Z
From: Duncan Murdoch
Subject: how to prove sum of probabilities to be one in R?
In-Reply-To: <CAGR+MS5nfrmsRqSsH1QUe2sZHw2X5ZtytqTeD5ZCMkS9yM4FFg@mail.gmail.com>
On 20/08/2020 3:42 a.m., Shaami wrote:
> Hi Dear
>
> I am facing a floating-point problem related to the sum of probabilities.
> It is really difficult to prove that sum of probabilities is 1 because of
> some minor differences. The MWE is as follows.
>
>> p1=0.99999999
>> p2=0.00000000003> p1+p2==1[1] FALSE
>
> The sum of probabilities is approximately 1. The difference from 1 is
> 1-p1-p2 = 9.97e-09, that is very small. I need to apply the sum of
> probabilities conditions in my many functions. But execution is halted
> because of floating point.
>
> Could anyone please guide about that?
Use the approximate test
isTRUE(all.equal(p1+p2, 1))
If the default tolerance of sqrt(.Machine$double.eps) (about 1.5e-8) is
wrong, change it:
isTRUE(all.equal(p1+p2, 1, tolerance = 0.1))
This says anything between 0.9 and 1.1 is equal to 1.
Duncan Murdoch